Skip to content

Latest commit

 

History

History
19 lines (10 loc) · 1.05 KB

README.md

File metadata and controls

19 lines (10 loc) · 1.05 KB

PRELIMINARY DATA -- IN DEVELOPMENT

WHAT

Graph-Interpreter is a scheduler of DSP/ML nanoApps designed with three objectives:

  1. Accelerate time to market

Graph-Interpreter helps system integrators and OEM who develop complex DSP/ML stream processing. It allows going fast from prototypes validated on a computer to the final tuning steps on production boards, by updating a graph of computing nodes and their coefficients without device recompilation.

  1. NanoApps repositories

It provides an opaque interface of the platform memory hierarchy to the computing nodes. It arranges the data flow is translated to the desired formats of each node. It prepares the conditions where nodes will be delivered from a Store.

  1. Portability, scalability

Use the same stream-based processing methodology for devices using 1 Kbytes of internal RAM up to multiprocessor heterogeneous architectures. Nodes can be produced in any programming languages. The Graph are portable when interpreted on another platform.

README